Shared machine learning

Patent metadata

Field	Value
Publication number	US-10445661-B2
Application number	US-201715717796-A
Country	US
Kind code	B2
Filing date	Sep 27, 2017
Priority date	May 5, 2017
Publication date	Oct 15, 2019
Grant date	Oct 15, 2019

Abstract

A network system may include a plurality of trainer devices and a computing system disposed within a remote network management platform. The computing system may be configured to: receive, from a client device of a managed network, information indicating (i) training data that is to be used as basis for generating a machine learning (ML) model and (ii) a target variable to be predicted using the ML model; transmit an ML training request for reception by one of the plurality of trainer devices; provide the training data to a particular trainer device executing a particular ML trainer process that is serving the ML training request; receive, from the particular trainer device, the ML model that is generated based on the provided training data and according to the particular ML trainer process; predict the target variable using the ML model; and transmit, to the client device, information indicating the target variable.

First claim

What is claimed is:

1. A network system comprising:
a computing system disposed within a remote network management platform, the remote network management platform comprising a scheduler device and a plurality of trainer devices, wherein each trainer device is configured to execute one or more machine learning (ML) trainer processes, wherein the computing system is configured to:
receive information comprising:
training data associated with the computing system, wherein the training data is to be used as basis for generating an ML model; and
a target variable to be predicted using the ML model;
wherein the information is received from a client device of a managed network, and wherein the remote network management platform remotely manages the managed network;

based on the received information, transmit an ML training request to the scheduler device, wherein the scheduler device assigns the ML training request to a particular trainer device of the plurality of trainer devices to execute a particular ML trainer process of the plurality of ML trainer processes, wherein the scheduler device provides an instance identifier of the particular computational instance to the particular trainer device, wherein the instance identifier enables direct communication between the particular trainer device and the particular computational instance;
receive a secure identifier from the particular trainer device;
verify that the particular trainer device is permitted to obtain the training data;
in response to verifying that the particular trainer device is permitted to obtain the training data, provide the training data to the particular trainer device executing the particular ML trainer process that is serving the ML training request;
receive, from the particular trainer device, the ML model that is generated based on the provided training data and according to the particular ML trainer process;
predict the target variable using the ML model; and
transmit, to the client device, information indicating the target variable.

2. The network system of claim 1 ,
wherein the scheduler device is configured to schedule service of ML training requests amongst the plurality of trainer devices.

3. The network system of claim 2 , wherein the scheduler device is configured to:
make a determination that a location of the particular trainer device is threshold close to a location of the computing system, and
wherein the scheduler device assigns the ML training request to the particular ML trainer process based at least on the determination.

4. The network system of claim 2 , wherein the scheduler device is configured to:
make a determination that the particular ML trainer process is available to serve the ML training request, and
wherein the scheduler device assigns the ML training request to the particular ML trainer process based at least on the determination.

5. The network system of claim 2 , wherein the computing system is a first computing system, wherein the ML training request is a first ML training request, wherein the particular trainer device is a first trainer device, and wherein the particular ML trainer process is a first ML trainer process, the network system further comprising:
a second computing system disposed within the remote network management platform,
wherein the scheduler device is further configured to:
receive, from the second computing system, a second ML training request for scheduling of the second ML training request; and
in response to receiving the second ML request, assign the second ML training request to a second ML trainer process, wherein assignment of the second ML training request to the second ML trainer process causes a second trainer device to execute the second ML trainer process serving the second ML training request.

6. The network system of claim 5 , wherein the second trainer device is different from the first trainer device, and wherein the second ML trainer process is different from the first ML trainer process.

7. The network system of claim 5 , wherein the first and second trainer devices are the same particular trainer device, and wherein the second ML trainer process is different from the first ML trainer process.

8. The network system of claim 5 , wherein the first and second trainer devices are the same particular trainer device, and wherein the first and second ML trainer processes are the same particular ML trainer process.

9. The network system of claim 8 , wherein the scheduler device is configured to:
determine that the particular ML trainer process is available after completing serving of the first ML training request, and
wherein assigning the second ML training request to the particular ML trainer process is further in response to determining that the particular ML trainer process is available after completing serving of the first ML training request.

10. The network system of claim 2 , wherein the information received from the client device specifies a training time, and wherein the scheduler device assigning the ML training request to the particular ML trainer process comprises the scheduler device assigning the particular ML trainer process to serve the ML training request at the specified training time.

11. The network system of claim 1 , wherein the computing system is configured to:
transmit a randomly generated bitstring along with the ML training request to one of the plurality of trainer devices;
wherein receiving the secure identifier from the particular trainer device comprises receiving the randomly generated bitstring from the particular trainer device when the particular trainer device requests that the computing system provide the training data; and
wherein verifying that the particular trainer device is permitted to obtain the training data comprises verifying that the randomly generated bitstring received from the particular trainer device is identical to the randomly generated bitstring transmitted by the computing system.

12. The network system of claim 1 , wherein the particular trainer device comprises a temporary data storage device, and wherein the particular trainer device is configured to:
store the training data at the temporary data storage device while the particular ML trainer process is serving the ML training request; and
delete the training data from the temporary data storage device after the particular ML trainer process completes the serving of the ML training request.

13. The network system of claim 1 , wherein the ML training request is a first ML training request, wherein the particular trainer device is a first trainer device, wherein the particular ML trainer process is a first ML trainer process, wherein the training data is first training data, wherein the target variable is a first target variable, wherein the ML model is a first ML model, wherein the received information comprises:
second training data associated with the computing system, wherein the second training data is to be used as basis for generating a second ML model; and
a second target variable to be predicted using the second ML model, and wherein the computing system is configured to:
transmit a second ML training request to one of the plurality of trainer devices, wherein the second ML training request is also based on the received information;
provide the second training data to a second trainer device executing a second ML trainer process that is serving the second ML training request;
receive, from the second trainer device, the second ML model that is generated based on the training data and according to the second ML trainer process;
predict the second target variable usin

Assignees

• Servicenow Inc

Inventors

• Bendre Nikhil
• Ros Fernando
• Govindarajan Kannan
• Jayaraman Baskar
• Thakur Aniruddha
• Palapudi Sriram
• Karakusoglu Firat

Classifications

• G06F9/46

  Multiprogramming arrangements · CPC title

• H04L67/00

  Network arrangements or protocols for supporting network services or applications (user-to-user messaging H04L51/00; network arrangements, protocols or services for supporting real-time applications in data packet communications networks H04L65/00) · CPC title

• H04L41/12

  Discovery or management of network topologies · CPC title

• G06N20/00 Primary

  Machine learning · CPC title

• H04L67/32

  Electricity · mapped topic